This proposal aims at improving methods of survival analysis and determining prognosis using SEER data. Specific topics to be studied include: robustness of parametric and non-parametric methods, small sample behavior of methods known to be good asymptotically, the effects of incomplete and delayed follow-up, comparing observed to expected survival curves, covariate adjustment in determining prognosis. Methods of procedure will include mathemical analysis. Monte Carlo studies, application to SEER data and comparison of these results to results from theoretical and Monte Carlo studies.